Latest KFLL download ... getting .NET runtime error

[robains]

I was actually testing to see if I could replicate this user's issue with KFLL over on AVSIM: http://forum.avsim.net/topic/471391-my-p3d-experience-was-ruined-2-hours-ago-net-problems-are-primary-suspects/

I was able to replicate with the latest downloaded KFLL ... I'm not using any Migration tool.  I re-installed the latest Add-On Manager also, but that didn't resolve the issue.

Event Log:

Log Name:      Application
Source:        Application Error
Date:          7/9/2015 12:27:18 AM
Event ID:      1000
Task Category: (100)
Level:         Error
Keywords:      Classic
User:          N/A
Computer:      Rob-5960X-PC
Description:
Faulting application name: Prepar3D.exe, version: 2.5.12946.0, time stamp: 0x555f1dfa
Faulting module name: unknown, version: 0.0.0.0, time stamp: 0x00000000
Exception code: 0xc0000005
Fault offset: 0xffb000a7
Faulting process id: 0x1c8
Faulting application start time: 0x01d0ba187389d4f8
Faulting application path: D:\Lockheed Martin\Prepar3D v2\Prepar3D.exe
Faulting module path: unknown
Report Id: ed5902e4-260b-11e5-9e41-54a050d5785d
Event Xml:
<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
  <System>
    <Provider Name="Application Error" />
    <EventID Qualifiers="0">1000</EventID>
    <Level>2</Level>
    <Task>100</Task>
    <Keywords>0x80000000000000</Keywords>
    <TimeCreated SystemTime="2015-07-09T07:27:18.000000000Z" />
    <EventRecordID>18822</EventRecordID>
    <Channel>Application</Channel>
    <Computer>Rob-5960X-PC</Computer>
    <Security />
  </System>
  <EventData>
    <Data>Prepar3D.exe</Data>
    <Data>2.5.12946.0</Data>
    <Data>555f1dfa</Data>
    <Data>unknown</Data>
    <Data>0.0.0.0</Data>
    <Data>00000000</Data>
    <Data>c0000005</Data>
    <Data>ffb000a7</Data>
    <Data>1c8</Data>
    <Data>01d0ba187389d4f8</Data>
    <Data>D:\Lockheed Martin\Prepar3D v2\Prepar3D.exe</Data>
    <Data>unknown</Data>
    <Data>ed5902e4-260b-11e5-9e41-54a050d5785d</Data>
  </EventData>
</Event>

Any hints of how we might be able to resolve this?

Cheers, Rob.

[virtuali]

None of our software use, is linked against to, communicates with, or is affected by the .NET runtimes in any way.  There are many 3rd party add-ons that use .NET, like UT2, REX, and many others, but none of our software use it.

The only think we ever made with .NET, was the Livery Manager for the QW757, but that's an external program that is launched separately from the sim, so it will surely not cause problems like crashes of the sim itself.

The Avsim forum is down at the moment, but if the .NET error you are referring to, is the "smartassembly" error, which is clearly a message from .NET, so it's not coming from any of our modules, it's a known problem that is caused by an UT2 issue. Since a problem with AI or an AI database at that location might appear on a specific scenery only, it's possible you might be mislead thinking it might be a scenery problem ( "because it only happens there..." ), when it's not.

[robains]

Thanks for the response Umberto.

I don't use UT2 but I do have MT6, only use "Soft Clouds" from REX and REX UI is .NET based but they just provide textures.

I'll keep digging, perhaps there is some conflict with another product ... if I disable all AI traffic (GA or Commercial) I don't get the error.

Cheers, Rob.

[virtuali]

if I disable all AI traffic (GA or Commercial) I don't get the error.

That's an important clue.

If you are sure your none of the tools you installed that might have something to do with AI are related to .NET (and should probably start with the sim as .DLLs, since .EXE files cannot crash it), you should check that all your AIs have correct (not corrupted and/or not missing) textures.

If this happened after installing KFLL, it can be easily mislead you thinking the scenery might have anything to do with it because, due to the different number of parkings and their airline preferences in the FSDT AFCAD compared to the default, will result in different AIs to be generated so, if the offending one only appears at FSDT KFLL, this explains why you only noticed the crash after its installation.

[robains]

Agree Umberto ... wasn't trying to point a finger at anyone.  I'm only starting here because it was the point of error that just happened to be after installing KFLL.

I have verified there are no duplicate AFCADs.  I've done the standard diagnostics by disabling all DLL.XML and EXE.XML entries that are NOT related to FSDT.  The other user with the same issue indicated they were not using any AI traffic products (I think), so that's interesting also.

My next step will be to uninstall FSDT KFLL and see if I still get the error at a default KFLL with AI enabled (MT6) ... if I do still get the error then I can certainly rule out FSDT KFLL as the problem.

I'll keep this thread updated ... I only started with FSDT as it was my initial point of error, hope you didn't take my post the wrong way, have no interest in blame, only in identification and finding resolutions.

Cheers, Rob.

[virtuali]

My next step will be to uninstall FSDT KFLL and see if I still get the error at a default KFLL with AI enabled (MT6) ... if I do still get the error then I can certainly rule out FSDT KFLL as the problem.

But if you do NOT get the error, it wouldn't mean KFLL was the problem either.

This because the default KFLL will generate both less AND different AIs, due to the difference of parking spots, both in number and in the airline preferences (the default AFCAD do not even use airline codes).

A more conclusive test would be, instead, trying to disable KFLL entirely, but copy its AFCAD in the Addon Scenery folder, to let it active.

This way, you should (with MT6) probably have the same set of AIs, but without the scenery. A crash in this case, would finally prove the scenery is not the problem, but some AI that was triggered to appear ONLY because of how the KFLL is made.

Of course, this wouldn't mean the KFLL AFCAD as a "problem", just that by pure chance, is made in a way that an AI that has a problem will be scheduled to appear there, and not on the default, because of all the differences in the number of parking spots, their radius and the airline codes.

Another potential area of investigation is: you said you ONLY get the error when you ( I assume you mean you were parked there ) were parked on Parking 102 GA. The location on the far north-east angle of the airport, and the fact the crash doesn't seem to happen  *immediately*, but after a while ( at least according to Dimon ), might even suggest that it's possible the problem is an AI scheduled to arrive from that direction, and maybe not even directed to KFLL (if it's a flying AI) or parked to *another* airport, which might be close enough to that position to be loaded.

[robains]

Agree it doesn't conclusive mean KFLL is the problem, just means it was the trigger.  As far as I know MT6 doesn't use .NET either ... the AI traffic is managed via .NET framework references in Prepar3d.exe.

Based on the error code returned which is indicative of referencing an element that doesn't exist or beyond the defined scope (like have a 10 item object collection and then referencing item(11)).

To clarify, with more testing I'd get the error no matter where my aircraft is, I could be flying around the airport and get the error. 

Is it possible there are just too many possible parking positions with the KFLL AFCAD which is triggering the AI traffic to have an issue?  Or perhaps the AI evaluated a parking score and attempt to use it but the space size was insufficient?  Or maybe a Radius overlap?  It it possible you could reduce the AI parking spots and provide me with some Test AFCAD files ?

I'll try your test tonight (in the office now) ... I see what you are getting at and seems like a good process.

Cheers, Rob.

[virtuali]

Is it possible there are just too many possible parking positions with the KFLL AFCAD which is triggering the AI traffic to have an issue?  Or perhaps the AI evaluated a parking score and attempt to use it but the space size was insufficient?  Or maybe a Radius overlap?  It it possible you could reduce the AI parking spots and provide me with some Test AFCAD files ?

If just the AFCAD itself was a problem, everybody should have it just the same and it should appear with default AI too. But as soon as you found that zeroing the AI density fixed it, I tried it, and replied on Avsim:

Of course, after Rob's findings the crash happens only at Parking 102 GA, I immediately tried it, with 100% density (commercial, GA and airport vehicles), and left the airplane sitting at that parking for about 1 hour. And of course, no problems whatsoever.

There's nothing special about the KFLL AFCAD. It has 127 parkings, compared to KLAX 206 and KJFK 215. Some of them overlap, but this happens in KJFK and KLAX too and all the rest of our airports and many other AFCADs out there, so this never used to be a problem.

Since I cannot replicate the crash with default AIs, It cannot just be an issue related to how the AFCAD is made, but it only appears in addition to something else, like a 3rd party traffic tool. Maybe P3D has issues when many AI are displayed or many AI are created/destroyed in a certain amount of time ?

KFLL is fairly close to another very large airport (KMIA), and the reality bubble is surely large enough to have AIs from both airports in memory (in addition to the ones in the smaller airports in the area), and two AFCAD for detailed airports so close, can generate a lot of traffic, if using real-life schedules.

[robains]

Good to know, I'll keep testing and see what more I can find.  Will also try your suggest test.

Cheers, Rob.

[robains]

Hi Umberto,

I re-installed KFLL and ran the test you suggested:

A more conclusive test would be, instead, trying to disable KFLL entirely, but copy its AFCAD in the Addon Scenery folder, to let it active.

This way, you should (with MT6) probably have the same set of AIs, but without the scenery. A crash in this case, would finally prove the scenery is not the problem, but some AI that was triggered to appear ONLY because of how the KFLL is made.

This time around I had NO CTDs ... in fact I had no CTDs in any situation??  So has me very puzzled ... made a quick video showing all the text variants with MT6:

The increased the video speed (NOT sim rate) 4X to reduce time duration of video ... but as you can see no CTDs in any variant including your test.  But more importantly no CTDs in the same test situation as before that did produce CTDs. 

So not really sure what to make of this?

Cheers, Rob.

[virtuali]

The increased the video speed (NOT sim rate) 4X to reduce time duration of video ... but as you can see no CTDs in any variant including your test.  But more importantly no CTDs in the same test situation as before that did produce CTDs.  

So, the crashes are just gone ? I really cannot explain it, maybe some file belonging to KFLL was corrupted locally on your system for some reason, and was restored by reinstalling it.

If there were some problem in the KFLL installer on in one of its file, we should have seen way more reports, instead of just two, considering the installer hasn't changed since February, and while the newest file in the set is the AFCAD (last modified in January 2015) most of the files belonging to the installation dates from 2012/2013.

[robains]

So far no more crashes ... the only variable is that I uninstalled and re-installed KFLL and ran those tests you suggested.  Exact same saved flight, time day, location, etc. etc.

Has me stumped ... as you can see form the video available VAS was very good too 1.34 GB free.

I've asked (multiple times) if LM can provide some type of debug/logging process initiated via command line to the prepar3d.exe ... no response so far.  It would go a LONG way in helping end user and 3PDs to resolve issues.  I've also asked for BGL validation tools so users know when a 3PDs have gone outside the scope of the SDK ... response was "maybe" which probably means "no".

It's understandable that products have CTD given the complexity of the environment, but it would be VERY useful to have some tools are our disposal that could make diagnostics more a "science" rather than guess work.  I'll keep fighting that fight with LM to get diagnostic tools, but given the nature of some 3PD's products, I'm sure their is some resistance to equalizing the playing field ... after all the SDK is for everyone that has purchased the main FS product.

Thanks again for you support.

Cheers, Rob.